We have developed a tumor model for studying the effects of aromatase inhibitors and antiestrogens in vivo. The model simulates postmenopausal breast cancer patients with estrogen-dependent tumors. This model utilizes human estrogen-dependent breast cancer cells transfected with the aromatase gene (MCF-7(CA)), which are inoculated in Matrigel s.c. into ovariectomized nude mice. The estrogen receptor-positive cells produce sufficient estrogen by aromatization to stimulate their proliferation and the formation of tumors in the mice. Using several different strategies, we have compared the aromatase inhibitors (letrozole and anastrozole) with antiestrogens (tamoxifen and fulvestrant). Aromatase inhibitors were more effective than tamoxifen and were more effective alone than when combined with antiestrogens. In addition, letrozole had a longer duration of effect than tamoxifen. The possibility of delaying the development of resistance to antiestrogens and aromatase inhibitors was investigated by alternating letrozole with tamoxifen every 4 weeks. However, although alternating treatment proved more effective than tamoxifen alone (tumors doubled in 16 weeks), tumors doubled in volume at about 18 weeks when treatment began with tamoxifen. When treatment started with letrozole and alternated with tamoxifen, tumors doubled in volume in about 22 weeks. However, tumors of mice treated with letrozole alone did not double in size until 35 weeks. These results demonstrate that this aromatase inhibitor is more effective and has a longer duration of response as a single agent than tamoxifen or in combination with tamoxifen.